characterization of the 2.4 ghz ism band electromagnet interference in a hospital environment

Upload: fernandosydor

Post on 04-Nov-2015

225 views

Category:

Documents


0 download

DESCRIPTION

Characterization of the 2.4 GHz ISM Band Electromagnet Interference in a Hospital Environment

TRANSCRIPT

  • Proceedings of the 25' Annual lniematmnal Conference of the IEEE EMBS Cancun, Mexico * September 17-21,2003

    Characterization of the 2.4 GHz ISM Band Electromagnetic Interference in a Hospital Environment

    Seshagiri Krishnamoorthy, Jeffery H. Reed, Christopher R. Anderson, P. Max Robert, and Srikathyayani. Srikanteswara

    Mobile and Portable Radio Research Group, Bradley Department of Electrical and Computer Engineering Virginia Polytechnic Institute and State University

    Blacksburg, VA 24061, USA

    Abstract- I n today's world, the ubiquitous nature of wireless personal systems has lead to .an increasing concern regarding the safety and reliability of the Electromagnetic Environment (EME) in hospitals and clinics. Additionally, modern hospitals a re becoming populated with wireless medical applications. Research conducted on the EME so far has only considered frequency ranges below 2 GHz. However, many short- range wireless devices such as Bluetooth and Wireless LANs operate in the 2.4 GHz ISM band and are already seeing significant use for medical applications such as transferring patient records. These devices are also expected to be integrated in consumer electronics, to the point where patients or visitors may unknowingly bring wireless transmitters into the hospital environment.

    Electromagnetic Interference (EMI) measurements in the 2.4 GHz ISM band were performed in two hospitals. These measurements are considered to be very first eflort to characterize the entire 2.4 GHz ISM band in hospitals. Emphasis was given to critical locations such as Emergency Rooms (ER), Intensive Care Units (ICU), Surgery blocks, and Radiology.

    Keywords-2.4 GHz ISM, Electromagnetic Interference (EMI), hospital environments,, Bluetooth, WLAN, microwave ovens

    I. INTRODUCTlON

    The 2.40-2.4835 GHz kequency band, also known as the ISM (Industrial Scientific and Medical) band, is a license free band allocated for a variety of consumer applications. While no license needs to he purchased to operate devices in-this band, regulations exist governing its use under Part 15 of the FCC code. Short-range wireless devices such as Bluetooth and Wireless LANs operate in the ISM band. Over the next 2 to 5 years, these devices'are expected to become widespread throughout hospital and medical enviromzknts for applications such as data transfer, patient monitoring and .inventory control. Also, as these wireless devices become more highly integrated in consumer electronics, there will be a number of unconscious personal wireless devices entering into and operating within hospitals. - '

    In this paper we look into the background EME that exists in a typical hospital environment in the 2.4 GHz ISM band. Currently, the FDA regulations recommend that non- life-supporting medical electrical equipment be resistant to background electric fields in the frequency range of 80 MHz

    to 2.5 GHz of 3 Vim (130 dBpVlm), increasing to 10 V/m (140 dBpV/m) for life-supporting medical equipment [I]. Understanding and analyzing the background interference is crucial, so that the medical community can be informed about the EM1 risk to life-support devices due to wireless devices operating io the ISM band. Secondly, knowledge of the existing Electromagnetic Environment (EME) is crucial for the development of wireless devices/applications that provide a high quality of service (QoS). MPRG has developed PRISM' (PRobe ISM), a software reconfigurable measurement device for studying the frequency spectrum of the ISM band.

    Measurements and characterization of EM1 in medical environments can be found in [ I - 81. These studies have been confined to spectrum below 2.0 GHz, primarily to study impact from mobile radios and cellular phones. So far there are no reported measurements or results on EM1 for the complete 2.4 GHz 1SM.band in hospitals, although Tan has performed EM1 measurements on a WLAN system operating at 2.42 GHz [9] and found it to be well below the FDA recommendations.

    11. MEASUREMENT SYSTEM

    PRISM (Fig. 1) is a fully automated' passive measurements system. PRISM consists of a vertical dipole sleeve antenna connected through a Low Noise Amplifier to a Hewlett-Packard (85948) spectrum analyzer. Finally, the spectrum analyzer is connected to the PC through a GPIB interface. The PRISM software was developed in LabVIEW, to take advantage of LabVIEWs powerful instrument control and data logging capabilities. Post processing of the data is performed in MATLAB. PRISM has an overall noise figure of 6 dB, enabling it to measure even extremely weak signals.

    I I Fig. I . System Model of PRISM.

    Beyond being a measuring system for the ISM band, PRISM can also be used to measure RF interference over non-contiguous frequency bands between 9 kHz and 2.9

    0-7803-7789-3/03/$17.00 0 2 0 0 3 IEEE 3245

  • GHz by using wideband &tennas. Additionally, PRISM can be configured to send data over the internet to enable monitoring EM1 kom a remote location.

    111. MEASUREMENT PLAN

    In order to characterize the EM1 in the 2.4 GHz ISM band, a measurement campaign was conducted at the Virginia-Maryland Regional College of Veterinary Medicine (VMRCVM) at Virginia Tech and the Carilion Roanoke Memorial Hospital (CRMH) in Roanoke, Virginia. Emphasis was given to locations such as Emergency Rooms, Intensive Care Units, Surgery blocks and Radiology. The medical devices and procedures used at the veterinary hospital and the regular hospital are similar in nature, but the density of equipment and personnel is higher in the human hospital.

    A. EM Measurements at the Virginia Tech Veterinary Hospital

    The VMRCVM is a fully equipped veterinary hospital located in the Virginia Tech central campus. The hospital facilities include a large animal clinic and a small animal clinic. A number of procedures including computed tomography (CT), Ultrasound, digital fluoroscopy, nuclear medicine and surgery are performed on .a regular basis. PRISM was setup at three different locations - Radiology, Clinical Pathology and the ECG lab.

    The first measurement site was located by the side of the lead shielded wall outside the radiology lab. The radiology office located adjacent to the measurement sice bad a turntable microwave oven. The second measurement site was located inside the clinical pathology lab. This lab bad a number of centrifuges using servo motors. The third measurement site was located in the hallway between the ECG, dentistry, special procedures, electrediagnostics lab, and anesthesiology facilities. The dentalicasting lab was equipped with an X-ray facility.

    PRISM was setup in a movable cart and the antenna was placed 0.95 meters above the ground. Measurements were recorded in two minute intervals for a period of one day outside the radiology unit, two days inside the clinical pathology lab and for 4 days outside the ECG block. Measurements were taken in 1 MHz bins, where the maximum signal power observed in each bin during a two second time window was recorded.

    E . EM1 Measurements in the Carilion Roanoke Memorial Hospital

    The Carilion Roanoke Memorial Hospital is designated as a level 1 trauma center and contains 500 patient beds. Apart kom the state-of-the-art medica1 facilities, the hospital has a sophisticated wireless infrastructure. Wireless medical telemetry systems are supported in the 560-614 MHz and

    700 MHz band. Wireless, phones are supported in the 900 MHz band and 802.11b WLANs are supported in the 2.4 GHz ISM band. Due to the extensive area of the hospital and the density of medical equipment and personnel, measurements were performed on three separ,ate floors (2, 4 and 7) which were expected to provide unique electromagnetic environments (floor 2 is used for emergimcy care, 4 is used for surgery and floor 7 is used for critical care). Five different measurement sites were selected on each of the three floors for recording EMI, based on surrounding areas where medical instruments were contained or procedures were performed-taking into account periodic activit:y and availability of space for performing the measurements without affecting patient care and the regular business of the hospital.

    The second floor of the Carilion Memorial Hospital accommodates the ER. This floor contains all the essential procedures and labs such as X-ray, Radiology (MRI), CT, ultrasound and nuclear modicine required for an emergency. The MRI rooms have walls embedded with copper and iron to provide RF shielding and magnetic shielding. The magnetic shielding is used to contain magnetic fields within the room to a safe level whereas the RF shielding is to minimize the level of WI coming kom outside the room and thereby providing a better signal to noise ratio for the MRI procedures. The X-ray rooms are protected by walls with lead shielding. The five measurement sites on this floor were chosen such that they could cover a variety of labs including MRI, CT, X-ray, ultrasound and the main ER room. Also, this section of the floor supports WLAN connectivity through an 802.11b Orinoco Access Point (AP) attached to the wall.

    The fourth floor ccsnsists of a number of surgical operating rooms and patient monitoring rooms. A considerable portion of this floor is designated as a clean environment and has reitricted access. The 4th floor also has a considerable amount of WLAN deployment and is supported by four 802.11b Orinoco WLAN APs. The five measurement sites were dhosen so as to record data as close as possible to the pat:ient monitoring rooms and the operating rooms. The operating rooms consist of a number of medical equipment including Electrosurgical Units, RF ablation, stereo tactic unit, surgical robots, anesthesia, ventilator, defibrillator, infusion pumps, CO2 laser, YAG laser, wireless phones and portable X-ray.

    The seventh floor of the Carilion Memorial Hospital consists of Intensive Care Units (ICU) as well as regular patient rooms. Equipment. used on this floor include patient monitors, defibrillator, telm monitor, wireless phone, infusion pumps, televisions, ventilators, humidifiers, sequential compression d.evice and bedside charting PCs. The seventh floor has the maximum utilization of WLAN connectivity in the hospital, including a number of wireless enabled carts used for providing medication for patients. There are five 802.11b Orinoco WLAN APs located along

    3246

  • the hallways and networking closets on this floor. The seventh floor also had pantries containing microwave ovens.

    In the Carilion measurement campaign, the ISM band was divided into 100 kHz bins and the maximum signal power observed in each bin during a two minute time window was recorded. On each of the fifteen sites, EM1 was recorded periodically every 5 minutes for a period of 24 hours on regular weekdays.

    Floor Blueprints which detail measurement locations for both hospitals can be found in [IO].

    I v . MEASUREMENT RESULTS

    Table 1 summarizes the maximum EM1 recorded at the three locations in both hospitals.

    In order to observe a worst case scenario on each floor of the Carilion Memorial Hospital, the recorded data from all five sites in a floor were time aligned and 3-D plots showing the maximum E-field recorded over all 5 sites as a function of frequency and time of day (i.e. maximum E-field recorded from any of the measurement sites at a particular time and frequency) were created. Fig. 2 and 3 shows the worst-case observed EM1 on 4th and 7th floors respectively. All individual plots for each measurement location at both the Veterinary Hospital and Carilion Hospital may be found in [IO].

    TABLE I Maximum Recorded EM1 from the Measurement Sites at the Virginia

    Tech Veterinary Hospital and the Carilion Memorial Hospital

    Radiology'

    ~ ~ ~

    Maximum Recorded EM1 Time ofDay _ID .V!- I ,,!- Observed Locatlo"

    ~ T 6:30pm ~ 102.9 0.139 C l i n i d Pathology' I 103.71 0.153 5:OO pm

    v . DISCUSSION OF RESULTS

    The measured field levels recorded at the Virginia Tech veterinary hospital were well helow the FDA recommended immunity level of 130 dBpV/m (3 V/m). The highest field strength observed was 103.71 dBpV/m (0.153 V h ) , with the majority of the EM1 activity at all three locations being observed during normal business hours. Most of the strong interference can be attributed to the use of microwave ovens

    Floor 2 (ER)' Floor 4 (Surgery)'

    Floor 7 (ICLJ)'

    in the hospital, which can be identified by the signal strength, spectral characteristics and frequency of operation [I11

    102.33 0.131 7:OO pm

    108.84 0.277 10:40 am

    114.83 0.552 2:20 am

    10

    2.4!

    t Fig. 1. Worst-case EM1 recorded from all 5 measurement sites on

    the 4' floor of the Carilion Roanoke Memorial Hospital.

    Fi 2. Worst-case EM1 recorded from all 5 measurement sites on the I' floor of the Carilion Roanoke Memorial Hospital.

    At the Carilion Memorial Hospital, pronounced levels of interference (60 - 100 dBpV/m) were observed at the five measurement sites of floor 2 between the hours of 7:OO am and 3:OO pm. There were three occurrences of high interference between 1 :OO am and 3:OO am which could be attributed to the use of microwave ovens. Microwave oven activity was also observed between the hours of 12:OO pm and 1:OO pm and around 7:OO pm. Two wideband signals appearing like a tunnel, centered at 2.412GHz and 2.437GHz were observed through out the day, due to the WLAN AF's operating on channel 1 and channel 6. The WLAN activity on channel 6 is expected because the South Section of the 2nd floor bas a "LAN AP operating on channel 6. The WLAN signal on channel 1 was caused by an AP operating in a neighboring floor; however, floor 1 and floor 3 have no WLAN deployment. Looking closer at the data, the channel 1 activity was only observed at single measurement site (out of the-five sites on the 2nd floor).

    3247

  • Because this site was located next to the elevators, i t is possible that the elevator shafts acted like waveguides and provided a path for WLAN signals to propagate from a higher floor. Looking at the Blueprint, only floor 7 had an AP on channel 1 located in a closet beside this group of elevators.

    On the 4 Floor, moderate EM1 was observed between the hours of 7:30 am and 11:OO pm. Increased activity (75 - 105 dBpV/m) was observed during lunch time between 11:OOam and 12:45 pm possibly due to use of microwave ovens, although the E-fields extend to the entire range of the ISM band. EM1 similar to microwave ovens was also measured around 6:OO pm, 8:OO pm, 1O:OO pm and 9:OO am. Though the 4th floor did have three active WLAN channels (four APs) deployed, strong signals (65 dBFV/m) only on channel 1 were observed. In fact, W A N channel 1 activity was observed from the data recorded f?om three sites, located close to the elevators. As discussed above, these signals should have propagated from the WLAN AP on the 7th floor

    Continuous activity throughout the day was observed on the 7th floor and pronounced levels of interference (90 - 110 dBpV/m) were observed at all five measurement sites. The 7th floor bas 6 WLAN A P s deployed along hallways and inside networking closets. The presence of strong interference from WLAN APs operating on channel 1, 6, IO and I 1 was observed. Additionally, microwave oven interference was observed between the hours of 12:OO pm - 1:OO pm, 9:OO pm - 1O:OO pm, 1:OO am - 2:OO am, and 5:OO - 5:30 am.

    VI. CONCLUSION

    An extensive measurement campaign at the Virginia-Maryland College of Veterinary Medicine at Virginia Tech and the Carilion Memorial Hospital in Roanoke, Virginia was undertaken to investigate the EM1 in the 2.4 GHZ ISM band. We have presented here what we believe to be the first study of EM1 in the 2.4 GHz ISM band in hospital environments. EMI as a function of frequency and time of occurrence at various locations including radiology, MRI, Critical Care Units, ECG, Clinical pathology and surgery units was recorded.

    The majority of EM1 activity at the Virginia Tech veterinary hospital was observed during normal business hours, and can be primarily attributed to the use of microwave ovens. In the Carilion Memorial Hospital, the EM1 recorded on floors 2, 4 and 7 were found to be time varying and dynamic. WLANs and Microwave ovens accounted for a majority of the recorded interference, though microwave oven activity was less prominent as compared to the veterinary hospital. The maximum interference levels observed at both the veterinary hospital and the Carilion memorial hospital are far below the recommended immunity level of 130 dBpV/m (3 V/m) for medical devices specified by the FDA.

    The impact of wireless devices such as Bluetooth on the electromagnetic environment is also expected to be minimal, since the device output level (maximum of 20 dBm at the antenna input) is comparable to the recorded levels of EMI.

    ACKNOWLEDGMENT

    The authors would lik:e to thank Carilion for providing access to the Roanoke Memorial Hospital. The Carilion Biomedical Institute, the Medical Automation Research Center at the University ofVirginia and the MPRG affiliates program sponsored this work.

    REFERENCES

    Regulatory Compliance Analysis EMC Standard, IEC (EN)60601 -I -2 (2001 Edition). Boisvert, P., Segal, E., Pavlasek, T., Retfalvi, S., Sebe, A., and Caron, P., Preliminary Survey o f the Electromagnetic Interference in Hospitals, IEEE lntemational Symposium on Electromagnetic Compatibility. August 1991, pp. 214 -219. Boivin, W.S., Boyd, S.M., Coletta, J.N., Harris, C.D., and Neunaber, L.M., Measurement of Electromagnetic Field Strengths in Urban arid Suburban Hospital Operating Rooms, IEEE 19th Annual International Conference an Engineering in Medicine and Biology. November 1997, ~01.6. pp. 2539 -2542. Davis, D., Skulic, B.. Segal, B., Vlach, P., and Pavlasek, T., Hospital Emergency Room Electromagnetic Environment, IEEE 19th Annual lntemational Conference on Enaneering in Medicine and Biology, November 1997, 01.6, pp. 2543 -2546. Davis, D., Skulic, B., Segal, B.. Vlach, P., and Pavlasek, T.. Variation of Emerg.ency-Room Electromagnetic-Interference Potential. IEEE Intemational Symposium on Antennas and Propagation, lune 1998, ~01.4, pp. 1996-1999. Vlach, P., Segal, B., and Pavlasek, T., The Measured and Predicted Electromagnetic Environment at Urban Hospitals, IEEE International Symposium on Electromagnetic Compatibility, August 1995, pp. 4 -7. Vlach, P., LiuHinz, C:., Segal, B., Skulic, B., and Pavlasek, T., The Electcomametic :Environment due to Pomble Sources in a Typical Hospital Roam, IEEE 17th Annual Engineering in Medicine and Bialog,y Society Conference, September 1995. v d l , pp. 683 -684. Phaiboon, S.. and Somkuampanit, S., Modeling and Analysis the Effect of Radio-Frequency Fields in Hospitals to the Medical Equipment, IEEE TENCON 2000, v01.I. pp.92-95. Young, C., Ahmed Saoudy, S., and Budwill, S.. EM1 Levels at a Patient Care Localion in a Hospital. IEEE Canadian Conference on Electrical and Computer Engineering, May 1997, vol. 2. pp. 625-628. Tan, K-S., Hinberg, I., and Wadhwani. J., Electromagnetic Interference in Medical Devices: Health Canadas Past and Current Perspectives; and Activities, IEEE Inrernalionol Symposium on Elecrromognelic Comparibilily, 2001, vol.2, pp. 1283 -1288.

    [ I O ] Krishnamoorthy, S., Interference Measurements and Throughput Analysis for 2.4 GHz Wireless Devices in Hospital Environments, Masters Thesis,.Virginia Polytechnic Institute and Stare University, htlrr:llscholar.lib.vt.edunheses/index.html, April 2003.

    [ I I ] Krishnamoorthy. S., Robert, M., Srikanteswaa, S.. Valenti, M.C., Anderson, C.R., Reed, J.H., Channel Frame Error Rate for Bluetwth in the Presence of Minawave Ovens, lEEE Vehicular Technolog:, Conference. Fall 2002, V01.2, pp. 927 - 931.

    3248